n the one hand, we keep hearing that AI is about to turn e-commerce upside down. It will write product descriptions, analyse performance, set up campaigns, optimise the store and even tell us what to do to increase sales.
On the other hand, spend just one day with a team responsible for a large online store and you quickly realise how little this vision has to do with reality.
Running e-commerce is not just about clicking through an admin panel. It is a constant stream of decisions: what should take priority today, where margin is being lost, how to respond to declining conversion, how to manage inventory, and how to align marketing, sales, logistics, customer service, IT and the wider business.
Then there is seasonality, pressure on results, promotions, integrations, returns, stock availability issues and hundreds of smaller decisions made every single day.
That is why the question, “Will AI replace the e-commerce manager?” is not particularly interesting.
A much better question is:
Will the companies that learn to use AI faster and more intelligently gain an advantage by shortening the journey from data to decision, and then from decision to action?
Shopify Sidekick is one of the best examples of what this shift can look like in practice.
Sidekick is not ChatGPT added to the Shopify admin
Put simply, Shopify Sidekick is an AI assistant built directly into Shopify Admin. You can communicate with it using natural language, asking both for information and for specific actions to be carried out.
The key, however, is not AI itself, but the context in which it operates.
Sidekick has access to store data according to the permissions of the logged-in user and can perform actions directly within the admin. Shopify also emphasises that users can review and approve changes before they are applied.
This is fundamentally different from using a generic AI model in a separate browser tab.
Of course, you can ask ChatGPT or another tool to analyse a sales report. But first, you need to provide the data, explain its structure, describe the business context and then manually transfer the recommendations back into the system.
Sidekick is already where the work is actually being done.
1. Analytics without digging through endless reports
One of the most obvious use cases for Sidekick is working with data.
Instead of starting by opening Analytics, selecting reports, applying filters, exporting CSV files and building yet another spreadsheet, you can start with a question.
For example:
“Show me sales, costs, profit, returns and discounts for the 100 best-selling products.”
Based on this, Sidekick can generate a ShopifyQL query, prepare a data exploration, display the results as a chart and export the report.
This means you can ask questions that are much closer to everyday business language:
“Which products have lost the most sales over the last 30 days?”
“How has sales performance in this category changed month on month?”
“Which products should we restock first?”
Sidekick can also support demand forecasting and identify products that need replenishing. Shopify even highlights the ability to create purchase orders with Sidekick.
That does not mean AI suddenly becomes your e-commerce director.
If a product’s conversion rate drops by 20%, Sidekick can help identify the decline. What it may not know is that three days earlier your biggest competitor reduced its price by 25%, an influencer criticised the product on TikTok, or your supplier has been delaying shipments for two weeks.
AI accelerates analysis. People still need to provide the business context.
That is the right way to think about this capability.
2. Shopify Flow automations described in natural language
The second area where Sidekick can save much more than just a few minutes is Shopify Flow.
Flow makes it possible to automate store processes using triggers, conditions and actions. The problem is that as the number of automations grows, configuring them becomes another area that requires both time and a good understanding of the system.
Sidekick changes how this work can begin.
Instead of building a workflow from scratch, you can describe the desired process in natural language, for example:
“If a customer spends more than PLN 1,000 in the store, add the VIP tag.”
Or:
“Tag orders containing fewer than three products as ‘small’.”
Based on that instruction, Sidekick can build the relevant workflow in Shopify Flow. The user can then review it, ask AI to make adjustments or edit it manually.
There is, however, an important caveat.
Sidekick should not be treated as an “automate my business” button.
A workflow generated by AI is inactive by default. It needs to be reviewed, tested and only then enabled manually. Shopify explicitly warns that in more complex processes, Sidekick may not account for specific business logic.
This is a very sensible approach.
AI is excellent at translating human intent into the first version of a mechanism. Responsibility for whether that mechanism actually does exactly what the business expects still remains with us.
3. Less time spent asking: “Where is that setting?”
Anyone who has worked with a sophisticated e-commerce platform knows this problem.
We know what we want to do. We just cannot remember where the relevant setting is.
And suddenly, a five-minute change turns into a search through documentation, Google results and multiple tabs in the admin panel.
Sidekick works here almost like a context-aware Shopify manual.
It can navigate to the right place in the admin, fill in forms and support tasks such as creating and editing products, collections, discounts or customers. In B2B stores, it can also help create companies. It can also propose changes to existing orders, which the user can then approve.
In practice, this changes the way we interact with the system.
Instead of thinking:
“Where has Shopify hidden setting X?”
we can increasingly start with:
“I want to achieve X. Help me do it.”
It may seem like a small difference. But across hundreds of interactions every month, it can have a very tangible impact on team productivity.
4. Content and merchandising: much more than product descriptions
Text generation is probably the least surprising Sidekick capability.
It can help create product descriptions, marketing copy, blog posts and other materials. It can also generate images, customer segments and discounts.
But generating copy alone is hardly revolutionary. Almost every AI model can do that today.
The context is far more important.
If AI operates directly inside the system, it can help you move from:
“I need a promotion for this group of products”
to preparing a segment, discount, content or relevant campaign elements without repeatedly moving information between tools.
The same applies to working on the storefront.
Within the theme editor, Sidekick can modify existing section and block settings, such as typography, spacing, colours or other design parameters. The changes still need to be saved manually. So this does not mean AI can design an entirely new storefront from scratch based on a single sentence.
What it does offer is a much faster route to the first version of a change.
5. Sidekick is starting to move beyond Shopify Admin itself
An even more interesting direction for Sidekick can be seen in the platform’s latest capabilities.
In Shopify Editions Spring ’26, Shopify announced Sidekick integrations with selected apps, including Judge.me, Klaviyo, Loop and Smile. This means the assistant can not only answer questions about data held directly in Shopify, but also work with supported applications.
That matters because a real e-commerce stack rarely ends with the commerce platform itself.
An even more interesting example is apps generated by Sidekick.
On Grow, Advanced and Shopify Plus plans, Sidekick can create a simple internal app based on a user’s description. Shopify gives examples such as a tool recommending products for reordering, a discount-link generator or a simple task tracker for the team.
This still does not mean development is becoming obsolete.
These apps have clear limitations: they work inside Shopify Admin and cannot simply create arbitrary storefront, checkout or external-system functionality.
They do, however, point towards a future in which the line between using software and building simple tools tailored to your own processes begins to blur.
6. From answering questions to suggesting what to do next
Until now, most interactions with AI have started with a person.
We have a question. We open a tool. We write a prompt.
Shopify is trying to take this a step further with Sidekick Pulse.
Pulse analyses store data and can proactively surface recommended actions in Shopify Home. Once a recommendation is selected, Sidekick prepares a list of next steps, but the user still needs to approve any changes. The feature is currently in early access and available only to selected merchants.
This could become one of the more important stages in the evolution of such tools.
AI stops being purely a search engine that requires us to ask the right question. It begins to act more like a system that proactively identifies areas requiring attention — although this is still far from autonomous business management.
How will Sidekick change the role of an e-commerce manager?
AI is not responsible for the P&L. It does not have to explain to the board why targets were missed. It does not negotiate budgets between departments. It does not understand every nuance of supplier relationships, competitive dynamics or the wider market situation.
What it can do is remove a significant amount of operational work from e-commerce teams.
Sidekick’s biggest limitation? Reality still matters
Technology cannot fix a poor business model.
If a store has a chaotic catalogue, incorrect product data, poorly thought-out pricing, inconsistent stock levels and misconfigured analytics, AI will not automatically solve any of those problems.
In fact, it may simply help you do the wrong things faster.
Ask it to generate hundreds of descriptions based on poor-quality data? You will get hundreds of poor-quality descriptions.
Analyse incorrectly configured data? You will get highly professional-sounding conclusions based on flawed assumptions.
There are also limitations to the context itself.
Sidekick can know a great deal about what is happening within the Shopify ecosystem and supported connected tools. That does not mean it has a complete view of the world outside them.
If sales of a jacket suddenly decline, the reason may be pricing, a missing size or deteriorating product-page conversion.
But it may also be an unusually warm winter.
AI sees the data. People need to understand the business.
That is why the most important principle for using Sidekick remains very simple:
AI prepares, analyses and recommends. People evaluate and take responsibility for the decision.
Shopify is building this principle into the product itself — Sidekick should not independently publish changes to the store without user approval.
Who can benefit from Sidekick the most?
For a small e-commerce business or a founder-led company, the biggest value lies in the breadth of support. One person who is simultaneously responsible for the offer, marketing, promotions and operations can navigate an increasingly complex system much more efficiently.
For a fast-growing brand, the benefit looks different. The problem is no longer a lack of expertise, but a lack of time. Sidekick can take over part of the work involved in analysis, configuration, preparing first drafts and automating repetitive tasks.
For large e-commerce organisations, the greatest potential may lie in process standardisation and scale. If a company carries out similar analyses or operations every week, Sidekick allows recurring prompts to be saved as so-called skills and run again using a shortcut. Shopify highlights recurring sales snapshots and forecasts as examples.
At the same time, the cost of an error increases with scale.
That is why, in larger organisations, the most important aspect of AI adoption will not be simply enabling the feature. It will be designing the right rules around control, data quality and accountability.
—
Want to use Shopify and AI for more than just generating content?
At Strix, we help brands design and develop Shopify solutions so that technology supports a specific business model, team processes and sales objectives.
If you are wondering where AI could genuinely reduce the workload of your e-commerce team, start with the process, not the tool.